1. Field
The present disclosure relates generally to an image processing method of extracting image data of an object from an image, and correcting the extracted image data, and an apparatus for performing the method.
2. Description of the Related Art
According to developments in image-capture technology, techniques for capturing an image of an object and storing a file of the captured image in storage media have become widespread. One technique has an effect of correcting a tilt of an object within an image. This allows an extracted image to be displayed on a display as if the extracted image is viewed from a point vertically above the object. The method renders an image of an object extracted from a perspective image by converting it into an image for a non-perspective plane.
A document may be also converted into digital data by using a scanner and the converted digital data may be stored, but the scanner is a large, generally non-portable apparatus. Accordingly, without using the scanner, digital data similar to data scanned by the scanner may be generated by capturing an image of a document by using a mobile device, which can capture images, and processing the captured image. However, in a process of separating only an object from a background of an image and extracting the separated object, inaccuracies arise. That is, accurately extracting the boundary of the object is a challenge in a case where the form of the object deviates from a rectangle. For instance, if the form is generally rectangular but a portion of the object is damaged or one or more vertices of the object does not lie at an intersection of two major sides (such as when a corner is cut out), accurate object separation becomes difficult.